Information retrieval is an important application for current computing systems, including search, content recommendation, and other personal assistant applications. Typical information retrieval systems focus on text for very specific information (e.g., location or text) that tags other information objects. Information retrieval for media such as images and videos typically relies on specific metadata inserted in or attached to the media manually by users. Media annotations may be made by users and augmented with simple tags such as time and location. Other image search experiences may rely on face recognition or color and image analysis techniques but may not consider higher-level contextual information associated with the image.
Users are increasingly producing, consuming, and sharing large amounts of digital media. For example, the spread of smartphones and other mobile devices has allowed many users to have constant access to a digital camera and other media capture capabilities. In light of the increasing volume of digital media available, many users may have difficulty retrieving or otherwise accessing images or other media after creation.